Experiencing power over AI: The fit effect of perceived power and desire for power on consumers' choice for voice shopping

Abstract The question of how artificial intelligence (AI) systems shape consumer behavior has attracted strong interest from HCI scholars and business practitioners. The current study focuses on AI assistants (e.g., Amazon Alexa) and a novel behavior – shopping via AI assistants – called voice shopping. The voice interaction between users and AI assistants features an unbalanced mode wherein users command their AI assistants to perform various tasks. We examine how the power experience in human-AI interaction impacts voice shopping. Our two-wave longitudinal data reveal that perceived power over AI assistants can reduce risk perception regarding voice shopping. The results of polynomial modeling and response surface analysis show that when perceived power fits desire for power, users perceive less risk and are more willing to shop via AI assistants. Our findings suggest that developers should be attentive to the match between power experience and desire for power when developing voice shopping functionalities. Our study also indicates that users could be vulnerable to the manipulation of the power experience by AI companies, which is of critical importance for policy-makers.